In recent years, many more users demand fast data transmission in a radio communication system as the volume of communication increases. The multicarrier transmission represented by OFDM (Orthogonal Frequency Division Multiplexing) gets attention as a way of communication to realize the fast and high-volume data transmission. The OFDM, which is used in IEEE 802.11a being a radio system of a 5 GHz-band or digital terrestrial broadcast, provides for simultaneous communication by arranging tens to thousands of carriers in a minimum frequency interval that does not induce interference theoretically. Generally in OFDM, these carriers are referred to as subcarriers which are digitally modulated with PSK, QAM or the like for communication. It is known that OFDM and forward error correction are combined to obtain strong tolerance to frequency selective fading.
The configuration of a data packet according to the IEEE 802.11a will be described with reference to FIG. 1. As shown in FIG. 1, a data packet used in the IEEE 802.11a consists of preambles A and B and a data signal. A preamble A is used for OFDM symbol synchronization and frequency synchronization, while a preamble B is used for identification of an antenna and estimation of a channel response. The two preambles are both predetermined signals, being signals also known to a receiving side.
FIGS. 12 and 13 show configuration examples of an OFDM modulating circuit and an OFDM demodulating circuit, respectively. In the drawings, the number of subcarriers in use is defined as N.
FIG. 12 is a functional block diagram of a usual OFDM modulation circuit. In FIG. 12, reference numeral 1000 denotes an forward error correction coding unit, reference numeral 1001 denotes a serial/parallel converting unit (S/P converting unit), reference numeral 1002 denotes a mapping unit, reference numeral 1003 denotes an IDFT (Inverse Discrete Fourier Transform) unit, reference numeral 1004 denotes a parallel/serial (P/S converting unit), reference numeral 1005 denotes a preamble A generating unit, reference numeral 1006 denotes a preamble B generating unit, reference numeral 1007 denotes a multiplexing unit, reference numeral 1008 denotes a guard interval inserting unit, reference numeral 1009 denotes a digital/analog converting unit (D/A converting unit), reference numeral 1010 denotes a radio transmitting unit and reference numeral 1011 denotes an antenna.
Transmitted information data is encoded in the forward error correction coding unit 1000. Then, the S/P converting unit 1001 performs serial/parallel conversion on the data by a data amount needed to modulate each carrier. The mapping unit 1002 modulates each carrier. Afterward, the IDFT unit 1003 performs IDFT. Although an example of using IDFT for OFDM modulation is illustrated herein, a general circuit often defines the number of points in a format 2n and uses the fast inverse Fourier transform (IFFT). In order to generate an OFDM signal of an N wave, a value 2n not less than N and nearest to N is generally used as the number of points of IFFT.
After the IDFT, the P/S converting unit 1004 converts the data into serial data, and then the multiplexing unit 1007 time-multiplexes the data with the preambles A and B, resulting in the packet configuration shown in FIG. 1. Then, the GI (guard interval) inserting unit 1008 inserts a guard interval. A guard interval is inserted to reduce interference between symbols in receiving an OFDM signal. Further, the data is converted into an analog signal in the D/A converting unit 1009 and then converted into a transmission frequency in the radio transmitting unit 1010, and finally a packet is transmitted from the antenna 1011.
FIG. 13 is a functional block diagram showing a configuration example of an OFDM demodulating circuit. As shown in FIG. 13, a receiver conducts reverse processing of the transmission in principle. In FIG. 13, reference numeral 1020 denotes an antenna, reference numeral 1021 denotes a radio receiving unit, reference numeral 1022 denotes an A/D (analog/digital) converting unit, reference numeral 1023 denotes a synchronizing unit, reference numeral 1024 denotes a GI removing unit, reference numeral 1025 denotes an S/P converting unit, reference numeral 1026 denotes a DFT (Discrete Fourier Transform) unit, reference numeral 1027 denotes a changeover switch, reference numeral 1028 denotes a preamble multiplying unit, reference numerals 1029 and 1030 denote multiplying units, reference numeral 1031 denotes a demapping unit, reference numeral 1032 denotes a P/S converting unit and reference numeral 1033 denotes an forward error correction decoding unit. However, also a demodulating circuit often uses FFT instead of DFT, as described above.
An electric wave received in the antenna unit 1020 is frequency-converted into a frequency band in which A/D conversion is possible in the radio receiving unit 1021. The A/D converting unit 1022 converts the data into a digital signal, for which the synchronizing unit 1023 conducts OFDM symbol synchronization using the preamble A. The GI removing unit 1024 removes a guard interval from the data. Afterward, the S/P converting unit 1025 performs serial/parallel conversion on the data. Then, the DFT unit 1026 performs DFT on the data, the changeover switch 1027 transmits the received preamble B subjected to DFT to the preamble multiplying unit 1028 and transmits the received data signal subjected to DFT to the multiplying unit 1029. The preamble multiplying unit 1028 multiplies (multiplies in a frequency domain) a complex conjugate of the received preamble B and the preamble B used in a transmitting side to estimate a channel response. In the following, estimation of a channel response using a preamble (preamble B) being a known signal and compensation of a channel response will be briefly described using numerical expressions. First, a preamble used in a transmitting side is represented by p(f) and an information signal is represented by s(f). Those are expressed as frequency domain signals herein. Additionally, after transmission of a preamble or an information signal, if channel response is defined as c(f), a received preamble p′(f) and a received information signal s′(f) are represented by the following equations. In the equations, c(f) is a complex function to give different amplitude and phase rotation for each subcarrier.Equation 1p′(f)=c(f)×p(f)  (1)s′(f)=c(f)×s(f)  (2)
However, thermal noise in a receiver is not considered herein for simplicity. For the receive signals, first, a complex conjugate of p′(f) is obtained in the preamble multiplying unit 1028, and the conjugate is multiplied by the preamble p(f) used in the transmitting side, as described previously. This multiplication is represented in the equation (3):
                    Equation        ⁢                                  ⁢        2                                                                                                                                                      p                                          ′                      *                                                        ⁡                                      (                    f                    )                                                  ×                                  p                  ⁡                                      (                    f                    )                                                              =                                                                    c                    *                                    ⁡                                      (                    f                    )                                                  ×                                                      p                    *                                    ⁡                                      (                    f                    )                                                  ×                                  p                  ⁡                                      (                    f                    )                                                                                                                          =                                                                    c                    *                                    ⁡                                      (                    f                    )                                                  ×                                                                                                p                      ⁡                                              (                        f                        )                                                                                                  2                                                                                        (        3        )            
The output (equation (3)) of the preamble multiplying unit 1028 is transmitted to the multiplying units 1029 and 1030, which multiply the output by a received data signal and a received preamble, respectively. An output of the multiplying unit 1029 is shown in the equation (4) and an output of the multiplying unit 1030 is shown in the equation (5):
                    Equation        ⁢                                  ⁢        3                                                                                                                                                      s                    ′                                    ⁡                                      (                    f                    )                                                  ×                                                      c                    *                                    ⁡                                      (                    f                    )                                                  ×                                                                                                p                      ⁡                                              (                        f                        )                                                                                                  2                                            =                                                c                  ⁡                                      (                    f                    )                                                  ×                                                      c                    *                                    ⁡                                      (                    f                    )                                                  ×                                  s                  ⁡                                      (                    f                    )                                                  ×                                                                                                p                      ⁡                                              (                        f                        )                                                                                                  2                                                                                                        =                                                                                                              c                      ⁡                                              (                        f                        )                                                                                                  2                                ⁢                                                                                                p                      ⁡                                              (                        f                        )                                                                                                  2                                ×                                  s                  ⁡                                      (                    f                    )                                                                                                          (        4        )                                                                                                                          p                    ′                                    ⁡                                      (                    f                    )                                                  ×                                                      c                    *                                    ⁡                                      (                    f                    )                                                  ×                                                                                                p                      ⁡                                              (                        f                        )                                                                                                  2                                            =                                                c                  ⁡                                      (                    f                    )                                                  ×                                                      c                    *                                    ⁡                                      (                    f                    )                                                  ×                                  p                  ⁡                                      (                    f                    )                                                  ×                                                                                                p                      ⁡                                              (                        f                        )                                                                                                  2                                                                                                        =                                                                                                              c                      ⁡                                              (                        f                        )                                                                                                  2                                ⁢                                                                                                p                      ⁡                                              (                        f                        )                                                                                                  2                                ×                                  p                  ⁡                                      (                    f                    )                                                                                                          (        5        )            
As shown in the equation (4), a received information signal is multiplied by an output of the preamble multiplying unit 1028, whereby influence of phase rotation by channel response c(f) is compensated and a signal having a phase equal to a transmitted signal s(f) is obtained. Then, the outputs (equations (4) and (5)) of the multiplying units 1029 and 1030 obtained in this way are inputted to the demapping unit 1031. A preamble subjected to channel response compensation in the equation (5) is used as a criterion to demap an information signal for each subcarrier. Then, the P/S converting unit 1032 serializes necessary data, the forward error correction decoding unit 1033 decodes the transmitted data.
One of examples of aiming fast and high-quality OFDM includes the way disclosed in the non-patent literature 1. Generally, different information bits are assigned to OFDM subcarriers. However, according to the non-patent literature 1, an identical information bit is assign to all subcarriers. In order to assign an identical information bit to all subcarriers in this way and keep a high transmission rate, the non-patent literature 1 proposes to set a different amount of phase rotation for each information bit and give the phase rotation being set to subcarriers, thereby enabling to assign different information bits to an identical subcarrier for transmission.
FIG. 14 shows a part of transmitter configuration disclosed in the non-patent literature 1. As shown in FIG. 14, in a transmitting device according to the non-patent literature 1, an information bit (for BPSK modulation in the non-patent literature 1) mapped by a mapping unit 1050 is copied by the number of subcarriers (the number of subcarriers is N herein) by copy units 1051 and inputted to subcarrier demodulating and phase rotating units 1052. The subcarrier demodulating and phase rotating units 1052 assign information bits to all subcarriers and give phase rotation being set for each information bit to each subcarrier, as shown in FIG. 14. At that time, continuous phase rotation for adjacent subcarriers is given such that phase rotation given to the first subcarrier of a k-th information bit is 0, while phase rotation given to an n-th subcarrier is (n−1)Δθk. According to the non-patent literature 1, all of such phase rotation applied subcarriers are added, and outputs of the subcarrier modulating units and the phase rotating units for all information bits are further added in an adder 1053. A receiving device multiplies a complex conjugate of phase rotation given in a transmitting device, thereby compensating the phase rotation and restoring information data. The non-patent literature 1 discloses that it is possible to improve receiving features and ensure a high transmission rate by such configuration, compared to general OFDM.
[Non Patent Literature 1] D. A. Wiegandt, Z. Wu, C. R. Nassar, “High-throughput, high-performance OFDM via pseudo-orthogonal carrier interferometry spreading codes”, IEEE Transactions on Communications, vol. 51, no. 7, July 2003, pp. 1123-1134.
If a plurality of antennas simultaneously transmit different multicarrier signals, or if a terminal being positioned around a cell edge receives downlink transmission in an OFDM cellular system in which adjacent cells use an identical frequency band, a plurality of different multicarrier signals are mixed in a receiving side, so that the respective signals interfere with each other. In such a case, it is very difficult to identify which antenna has transmitted a received signal or which base station has transmitted the signal. Because of this, there has been a problem in that the accuracy of estimation of a channel response deteriorates significantly.
It is an object of the present invention to improve the accuracy of estimation of channel responses in receiving signals from a plurality of antennas.